28,548 claims · page 229 of 286
| Verdict | Person | Kind | Said | Claim | Topic |
|---|---|---|---|---|---|
| Not researched | Nate Silver | Factual | Nov 2016 | The error in national polls in the 2016 presidential election was no worse than usual. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | Polls tend to replicate one another’s mistakes Nate Silver · Not checkable · Nov 2016 | |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | So there wasn’t any reasonable way to construct a polling-based model that showed Trump ahead. Nate Silver · Not checkable · Nov 2016 | |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | This isn’t just a case of hindsight bias. Nate Silver · Not checkable · Nov 2016 | |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | We think the outcome — and particularly the fact that Trump won the Electoral College while losing the popular vote — validates important… Nate Silver · Not checkable · Nov 2016 | |
| Resolved true | Nate Silver | Factual | Nov 2016 | FiveThirtyEight's final 2016 forecast, issued early on election Tuesday evening, gave Trump a 29 percent chance of winning the Electoral College. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not researched | Nate Silver | Factual | Nov 2016 | For most of the 2016 presidential campaign, FiveThirtyEight's forecast gave Trump much better odds of winning than other polling-based models. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | We strongly disagree with the idea that there was a massive polling error. Nate Silver · Not checkable · Nov 2016 | |
| Mostly true | Nate Silver | Factual | Nov 2016 | The main reason FiveThirtyEight's 2016 model gave Trump a better chance than other models was its assumption that polling errors are correlated. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not researched | Nate Silver | Factual | Nov 2016 | Historically, polling error in US presidential elections has been larger in state polls than in national polls. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | But it shouldn’t have been that much of a surprise based on the polls Nate Silver · Not checkable · Nov 2016 | |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | one of the reasons to build a model — perhaps the most important reason — is to measure uncertainty and to account for risk. Nate Silver · Not checkable · Nov 2016 | |
| Unresolvable | Nate Silver | Factual | Nov 2016 | Inferences from early voting data have a fairly bad track record in predicting election outcomes, including in 2016. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | Because the outcomes in these Midwestern states were highly correlated, having problems in any one of them would mean that Clinton was… Nate Silver · Not checkable · Nov 2016 | |
| Not researched | Nate Silver | Factual | Nov 2016 | Pre-election 2016 polls showed a fairly close and highly uncertain presidential race. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Unresolvable | Nate Silver | Factual | Nov 2016 | Most mainstream journalists would have given Trump much lower odds than the 30 percent chance FiveThirtyEight gave him. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | If polling were perfect, you wouldn’t need to do this. Nate Silver · Not checkable · Nov 2016 | |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | people mistake having a large volume of polling data for eliminating uncertainty. It doesn’t work that way. Nate Silver · Not checkable · Nov 2016 | |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | Clinton’s Electoral College leads weren’t very robust. Nate Silver · Not checkable · Nov 2016 | |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | But as Election Day approaches, we think models ought to be fairly aggressive about detecting polling movement. Nate Silver · Not checkable · Nov 2016 | |
| Mostly true | Nate Silver | Factual | Nov 2016 | Clinton significantly underperformed her polls across the Midwest and Rust Belt in the 2016 election, costing her the Electoral College. Nate Silver · Factual · Nov 2016 | AI timelines and forecasting |
| Not checkable | Nate Silver | Not checkable | Nov 2016 | There was widespread complacency about Clinton’s chances in a way that wasn’t justified by a careful analysis of the data and the… Nate Silver · Not checkable · Nov 2016 | |
| Not researched | Nate Silver | Factual | 11 Oct 2016 | FiveThirtyEight's ratings rank the 2016 Chicago Cubs among the 25 or 30 best regular-season teams in history. Nate Silver · Factual · 11 Oct 2016 | Other |
| Resolved true | Nate Silver | Factual | 11 Oct 2016 | FiveThirtyEight's forecast of Clinton's chances is based on objective analysis of polling in presidential elections dating back to 1972. Nate Silver · Factual · 11 Oct 2016 | Science and research methods |
| Unresolvable | Nate Silver | Factual | 11 Oct 2016 | FiveThirtyEight's forecast gave Trump only a 20 percent chance of winning given his polling numbers at the time of the second debate. Nate Silver · Factual · 11 Oct 2016 | Other |
| Not researched | Nate Silver | Factual | 11 Oct 2016 | As of October 2016, the Chicago Cubs would have to win three playoff series in a row to win the World Series. Nate Silver · Factual · 11 Oct 2016 | Other |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | There are some people in that campaign that are sane enough to know those polls are B.S. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | It’s just a lot of things can go wrong when you have to win a bunch of short series. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | It’s based on data. It’s based on history. Nate Silver · Not checkable · 11 Oct 2016 | |
| Mostly true | Nate Silver | Factual | 11 Oct 2016 | The two scientific polls after the second 2016 presidential debate showed a Clinton win, but a more marginal one than after the first debate. Nate Silver · Factual · 11 Oct 2016 | Science and research methods |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | baseball is a random enough sport where — nevermind any curses or anything like that — it’s hard to win three series in a row. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | These are little surveys that are voluntary surveys. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | People can spam these surveys and they should not be mentioned in the same breath as scientific polls Nate Silver · Not checkable · 11 Oct 2016 | |
| Mostly true | Nate Silver | Factual | 11 Oct 2016 | Silver said that all the scientific polls after the first 2016 presidential debate showed Clinton had won by a pretty emphatic margin. Nate Silver · Factual · 11 Oct 2016 | Science and research methods |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | There is legitimate disagreement among polls. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | It’s a terrific baseball team. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not researched | Nate Silver | Factual | 11 Oct 2016 | Before the second 2016 presidential debate, Trump trailed Clinton by five or six points nationally, with large deficits in swing states. Nate Silver · Factual · 11 Oct 2016 | Other |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | So we’re not saying this Cubs team isn’t terrific. Nate Silver · Not checkable · 11 Oct 2016 | |
| Resolved true | Nate Silver | Factual | 11 Oct 2016 | Scientific post-debate polls showed voters thought Clinton did better in the second 2016 presidential debate. Nate Silver · Factual · 11 Oct 2016 | Science and research methods |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | His position is fairly dire and at best it seems like he squandered an opportunity to turn the race around. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | That’s healthy. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | It's winnable maybe. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | So it’s not me subjectively telling you that it’s 80 percent. Nate Silver · Not checkable · 11 Oct 2016 | |
| Unresolvable | Nate Silver | Factual | 11 Oct 2016 | Silver said he thought "we" had the team, then playing in the NLDS, at "a 35 percent chance or something" of winning the three series in a row it still needed, and that the figure was rising. Nate Silver · Factual · 11 Oct 2016 | Other |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | some of the worst two-week stretches I can remember in presidential campaign history. Nate Silver · Not checkable · 11 Oct 2016 | |
| Not checkable | Nate Silver | Not checkable | 11 Oct 2016 | Frankly, after the events of the weekend, I think 80 percent might be low instead of high — my subjective take. Nate Silver · Not checkable · 11 Oct 2016 | |
| Mixed | Nate Silver | Factual | 11 Oct 2016 | Other forecasts gave Trump even lower chances than the 20 percent in Silver's forecast with numbers like those: five or ten percent. Nate Silver · Factual · 11 Oct 2016 | Other |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | For a review/tutorial on energy-based models and structured prediction on top of neural nets (or other models) see this paper Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | It allows us to train a discriminator as an unsupervised “density estimator”, i.e. a contrast function that gives us a low value for data… Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | A particular instance of energy-based models is factor graphs (non-probabilistic graphical models). Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | To enable deep learning systems to reason, we need to modify them so that they don’t produce a single output Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | But perhaps more interestingly, the generator can be seen as parameterizing the complicated surface of real data: give it a vector Z, and… Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | There are papers where people do amazing things with this, like generating pictures of bedrooms, doing arithmetic on faces in the Z vector… Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | This discriminator has to develop a good internal representation of the data to solve this problem properly. Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | This is what energy-based models are designed to do: give you a score for each possible configuration of the variables to be inferred. Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | The most important one, in my opinion, is adversarial training (also called GAN for Generative Adversarial Networks). Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | The Deep Learning textbook by Goodfellow, Bengio and Courville Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | an overview paper in Nature by myself, Yoshua Bengio and Geoff Hinton with lots of pointers to the literature Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | the Coursera course on neural nets by Geoff Hinton (starting to be a bit dated). Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | Combining learning systems with factor graphs is known as “structured prediction” in machine learning. Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | It can then be used as a feature extractor for a classifier, for example. Yann LeCun · Not checkable · 1 Aug 2016 | |
| Not checkable | Yann LeCun | Not checkable | 1 Aug 2016 | This, and the variations that are now being proposed is the most interesting idea in the last 10 years in ML, in my opinion. Yann LeCun · Not checkable · 1 Aug 2016 | |
| Unresolvable | Yann LeCun | Promise | 28 Jul 2016 | LeCun will teach his NYU deep learning course again in the Spring of 2017. Yann LeCun · Promise · 28 Jul 2016 | Their own work or organisation |
| Resolved true | Yann LeCun | Factual | 28 Jul 2016 | There have been a number of recent works putting graphical models on top of ConvNets and training end to end. Yann LeCun · Factual · 28 Jul 2016 | AI capabilities |
| Unresolvable | Yann LeCun | Factual | 28 Jul 2016 | There have been many proposals to combine neural nets and structured prediction, going back to the early 1990s. Yann LeCun · Factual · 28 Jul 2016 | AI capabilities |
| Mostly true | Yann LeCun | Factual | 28 Jul 2016 | The check reading system LeCun and a colleague built at Bell Labs in the early 1990s used structured prediction on top of convolutional nets, called Graph Transformer Networks. Yann LeCun · Factual · 28 Jul 2016 | AI capabilities |
| Unresolvable | Yann LeCun | Factual | 28 Jul 2016 | Videos of LeCun's 2015 NYU deep learning course lectures were taken down for legal reasons, while the slides remain available. Yann LeCun · Factual · 28 Jul 2016 | Their own work or organisation |
| Resolved true | Yann LeCun | Factual | 28 Jul 2016 | Yann LeCun gave a recent series of 8 lectures on deep learning at Collège de France in Paris. Yann LeCun · Factual · 28 Jul 2016 | Their own work or organisation |
| Unresolvable | Yann LeCun | Factual | 28 Jul 2016 | LeCun's Collège de France lectures were taught in French and later dubbed in English. Yann LeCun · Factual · 28 Jul 2016 | Their own work or organisation |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | WELL, I DON'T THINK IT WAS EXACTLY DRIVEN BY THAT, BUT IT CERTAINLY HAS TO BE ADDRESSED. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | I HAPPEN TO WELCOME IT BECAUSE I THOUGHT THERE ACTUALLY WAS SOME VALUE TO SOCIETY TO HAVING "THE WASHINGTON POST" CLEARLY CONTROLLED BY THE… Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | SO IT WAS AN ATTEMPT TO GET GENERAL AGREEMENT, NOT TOTAL AGREEMENT ON EVERY POINT. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Unresolvable | Warren Buffett | Factual | 21 Jul 2016 | Berkshire has a provision that in any corporate transaction Class A stock cannot be treated differently from Class B stock. Warren Buffett · Factual · 21 Jul 2016 | Their own work or organisation |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | I ORIGINALLY WOULD HAVE PREFERRED BERKSHIRE HATHAWAY TO BE A PRIVATE COMPANY. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Unresolvable | Warren Buffett | Factual | 21 Jul 2016 | Buffett has given away about 40% of his Berkshire shares as of July 2016. Warren Buffett · Factual · 21 Jul 2016 | Their own work or organisation |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | THEY EITHER GET BOUGHT OR THEY GO BROKE. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Resolved true | Peter Thiel | Factual | 21 Jul 2016 | Thiel does not agree with every plan in the Republican Party's platform. Peter Thiel · Factual · 21 Jul 2016 | Their own work or organisation |
| Resolved true | Warren Buffett | Factual | 21 Jul 2016 | When Buffett gives away his Berkshire shares, he always converts the Class A voting stock into Class B stock first. Warren Buffett · Factual · 21 Jul 2016 | Their own work or organisation |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | BUT IF YOU ASKED ME THAT QUESTION 40 YEARS AGO, I WOULD HAVE SAID I'D RATHER BE PRIVATE. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | I DON'T THINK IT SHOULD BE MANDATORY. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | BUT I LIKE THE FACT THAT PEOPLE PUT THEIR TRUST IN US AND THAT WE TREAT THEM LIKE PARTNERS AND THEY FEEL LIKE PARTNERS. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | CORPORATE GOVERNANCE HAS EVOLVED OVER THE YEARS. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not researched | Warren Buffett | Factual | 21 Jul 2016 | Buffett said Jamie initiated the corporate governance principles effort, calling him and the others probably a little more than a year before July 2016. Warren Buffett · Factual · 21 Jul 2016 | Philosophy, ethics and society |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | IT'S CREATING A DIALOGUE. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | IT WAS TO GIVE ENCOURAGEMENT TO COMPANIES THAT REALLY FELT UNEASY ABOUT GIVING GUIDANCE TO PERHAPS HAVE A LITTLE MORE BACKBONE ABOUT IT. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Peter Thiel | Not checkable | 21 Jul 2016 | When Donald Trump asked us make America great again, he's not suggesting a return to the past. Peter Thiel · Not checkable · 21 Jul 2016 | |
| Not researched | Warren Buffett | Factual | 21 Jul 2016 | Buffett said he thought they had held three meetings, each lasting somewhere between three and four hours. Warren Buffett · Factual · 21 Jul 2016 | Philosophy, ethics and society |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | BUT I CAN'T THINK OF ANYTHING WE GOT IN ANY FISTFIGHTS OVER. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | I'VE SEEN GUIDANCE PRODUCE SOME BAD RESULTS. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | WE WELCOME OTHER VIEWS ON IT. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Not researched | Warren Buffett | Factual | 21 Jul 2016 | Berkshire Hathaway has a dual-class share structure. Warren Buffett · Factual · 21 Jul 2016 | Their own work or organisation |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | IT'S BEEN SORT OF A NATURAL EVOLUTION, I WOULD SAY. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Unresolvable | Peter Thiel | Factual | 21 Jul 2016 | Among candidates in the 2016 race, only Donald Trump is being honest about America's economic decline. Peter Thiel · Factual · 21 Jul 2016 | Economics and markets |
| Unresolvable | Warren Buffett | Factual | 21 Jul 2016 | Berkshire does not use consultants or advisory organizations to direct how it votes its shares. Warren Buffett · Factual · 21 Jul 2016 | Their own work or organisation |
| Not checkable | Peter Thiel | Not checkable | 21 Jul 2016 | He is a builder, and it's time to rebuild America. Peter Thiel · Not checkable · 21 Jul 2016 | |
| Not checkable | Peter Thiel | Not checkable | 21 Jul 2016 | But today, our government is broken. Peter Thiel · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | BUT NO, I BELIEVE THAT THE BOARD OF DIRECTORS IS THERE TO REPRESENT SHAREHOLDERS, THE OWNERS. Warren Buffett · Not checkable · 21 Jul 2016 | |
| Resolved true | Warren Buffett | Factual | 21 Jul 2016 | In certain countries there are various ways for workers to participate in company governance. Warren Buffett · Factual · 21 Jul 2016 | Philosophy, ethics and society |
| Not checkable | Peter Thiel | Not checkable | 21 Jul 2016 | Of course, every American has a unique identity. Peter Thiel · Not checkable · 21 Jul 2016 | |
| Not checkable | Warren Buffett | Not checkable | 21 Jul 2016 | THIS IS NOT SOMETHING WE'RE GOING TO BE PASSING ALONG AND HOPING WE GET HUNDREDS OF SIGNATURES OR ANYTHING OF THAT SORT. Warren Buffett · Not checkable · 21 Jul 2016 |
“the check reading system my colleague and I built at Bell Labs in the early 1990s used a form of structured prediction on top of convolutional nets that we called “Graph Transformer Networks”.”
Yann LeCun, Quora answer 28 July 2016 Yann LeCun Quora Session Overview (KDnuggets)
The check reading system LeCun and a colleague built at Bell Labs in the early 1990s used structured prediction on top of convolutional nets, called Graph Transformer Networks.
Mostly true True in substance, wrong in a detail.
The second judge said resolved true; the verdict shown is the first judge's.
What would count as true or false
Judge as of 2016-07-28. True: published sources (e.g. LeCun, Bottou, Bengio & Haffner 1998, 'Gradient-Based Learning Applied to Document Recognition', and related papers) show that the Bell Labs check-reading system developed in the early-to-mid 1990s combined convolutional nets with a structured-prediction layer called Graph Transformer Networks. Mostly true: as stated, except for a minor detail such as the dates (e.g. mid-1990s rather than early 1990s) or the number of collaborators. Mixed: the system used convnets, but GTNs were a separate or later development not used in it, or vice versa. Mostly false: the system existed but did not use structured prediction on convnets in the way described. False: no such Bell Labs check-reading system by LeCun existed.
Evidence
- Yann LeCun - A.M. Turing Award Laureate ACM A.M. Turing Award · 1 January 2018
“As well as developing the convolutional approach, LeCun pioneered its application in “graph transformer networks” to recognize printed and handwritten text. This was used in a widely deployed system to read numbers written on checks, produced in the early 1990s in collaboration with Bengio, Leon Bottou and Patrick Haffner.”
- leon.bottou.org Léon Bottou (personal site hosting paper: Bottou, Le Cun, Bengio, CVPR 1997) · 1997 Own record
“A complete check reading system based on these concept is described. The system uses convolutional neural network character recognizers, combined with global training techniques to provides record accuracy on business and personal checks.”
- Gradient-based Learning Applied to Document Recognition Proceedings of the IEEE · November 1998 Own record
“A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques.”
Reasoning
The ACM Turing Award biography (E1), an independent source, says LeCun pioneered the application of the convolutional approach in "graph transformer networks", which was used in a widely deployed check-reading system produced in the early 1990s. It dates that system to the early 1990s. The 1998 Proceedings of the IEEE paper (E3) describes a graph transformer network for reading bank cheques, and the 1997 CVPR paper (E2) describes a check-reading system using convolutional character recognizers with global training techniques. It uses convolutional character recognizers with global training, which is a structured-prediction layer on top of convnets. The core claim holds, but E1 credits the work to LeCun with three collaborators (Bengio, Bottou and Haffner), not one colleague, so the verdict is mostly true.
Cite this claim
Yann LeCun, “The check reading system LeCun and a colleague built at Bell Labs in the early 1990s used structured prediction on top of convolutional nets, called Graph Transformer Networks.” (said 28 July 2016). Verdict: mostly true. Reliability API, claim c_4f9a162fcb6d, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_4f9a162fcb6d
Verdicts are made automatically; how verdicts and scores are made. The same data is on Hugging Face under CC BY 4.0.